Central catadioptric camera calibration using line image and mirror contour

Area of Science:

• Computer Vision
• Robotics
• Optical Engineering

Background:

• Central catadioptric cameras offer a wide field of view but require complex calibration.
• Existing calibration methods can be computationally intensive or require specific calibration patterns.

Purpose of the Study:

• To develop a novel linear algorithm for calibrating central catadioptric cameras.
• To recover both intrinsic and mirror parameters efficiently.

Main Methods:

• Utilizing the properties of the common self-polar triangle between line images and mirror contours.
• Employing two pairs of circular point images from line images to constrain the Image of the Absolute Conic (IAC).
• Recovering camera parameters without prior principal point information.

Main Results:

• Demonstrated that a side of the common self-polar triangle passes through the principal point.
• Successfully constrained the IAC using limited circular point image data.
• Recovered intrinsic and mirror parameters accurately.

Conclusions:

• The proposed linear algorithm is effective and feasible for central catadioptric camera calibration.
• The method provides a robust approach for recovering camera parameters.
• Validated through experiments with both simulated and real-world data.